Supercomputing Asia 2025: MiTAC Computing Showcases Cutting-Edge AI and Servers

MiTAC Computing Technology Corp., a subsidiary of MiTAC Holdings Corp., will showcase its latest AI and HPC innovations at Supercomputing Asia 2025, in Singapore, taking place from March 11, at Booth B10.

The event highlights MiTAC’s commitment to delivering cutting-edge technology with the introduction of the G4520G6 AI server and the TN85-B8261 HPC server – both engineered to meet the growing demands of AI, ML, and HPC applications.

G4520G6 AI Server: Performance, Scalability, and Efficiency Redefined

The G4520G6 AI server redefines computing performance with an advanced architecture tailored for intensive workloads. Key features include:

Compute Power – Supports dual Xeon 6 Processors with TDP up to 350W, delivering high-performance multi-core processing for AI-driven applications.
Enhanced Memory Performance – Equipped with 32 DDR5 DIMM slots (16/ CPU) and 8 memory channels, supporting up to 8,192GB DDR5 RDIMM/3DS RDIMM at 6400MT/s for superior memory bandwidth.
GPU Support – Supports up to 8 full-height, dual-slot NVIDIA H200 NVL, NVIDIA H100 NVL, or NVIDIA L40S for scale-out AI training, ML, and rendering applications. Featuring the NVIDIA H200 GPU with 141GB of HBM3e memory and nearly 27 petaflops of FP8 deep learning compute, it accelerates AI and HPC workloads with exceptional speed and efficiency.
Superior Connectivity – Equipped with three standard PCIe 5.0 x 16 expansion slots and three OCP v3.0 mezzanine slots, the G4520G6 platform enables high-performance network connectivity, such as NVIDIA ConnectX NICs, BlueField-3 SuperNICs and DPUs, for clustered computing.
Advanced Cooling and Power Efficiency – Engineered for superior thermal management, it includes 8 hot-swap system fans and four external hot-swap fan modules. Four hot-swap CRPS power supplies provide (3+1) redundancy, delivering up to 9,600W with 80+ Titanium efficiency.

TN85-B8261: Versatile 2U Server for HPC and AI Workloads

The TN85-B8261 is a versatile 2U server designed to deliver performance for HPC and AI applications. Key highlights include:

Dual-Socket Performance – Supports AMD EPYC 9005 / 9004 processors, delivering computing power for AI and HPC workloads.
Scalable Memory Architecture – Features (12+12) DDR5 DIMM slots, supporting up to 6,144GB DDR5-6000 RDIMM/LRDIMM for high-speed data processing.
Flexible Expansion Capabilities – Supports up to 4 full-height, dual-slot enterprise H100 NVL and L40S GPUs for scalable AI training, ML, and rendering applications. Additionally, 2 half-height, half-length PCIe 5.0 x16 slots enable high-performance cluster networking deployment.
Optimized Connectivity and Management – Dual 10GBase-T LAN ports for high-bandwidth networking and an AST2600 BMC with IPMI 2.0 and Redfish support for remote server management.
Reliable Power Efficiency – (1+1) redundant 2,700W hot-swap CRPS power supplies with 80+ Titanium efficiency, delivering sustainable energy and operational reliability.

Driving Innovation in AI and HPC
The company is dedicated to pushing the boundaries of AI and HPC technology. The G4520G6 and TN85-B8261 servers are designed to empower enterprises and research institutions with compute power, scalability, and energy efficiency, ensuring optimal performance for next-gen workloads.